APT MRI Signature for Risk Stratification of Pediatric Medulloblastoma

Area of Science:

• Neuro-oncology
• Medical imaging
• Radiomics

Background:

• Medulloblastoma (MB) is a malignant brain tumor with four molecular subgroups (WNT, SHH, Group 3, Group 4) that have different prognoses.
• Accurate risk stratification is crucial for tailoring treatment strategies in pediatric MB.

Purpose of the Study:

• To develop and validate an Amino-Proton Transfer (APT) MRI-based radiomic signature for distinguishing between the molecular risk subgroups of pediatric medulloblastoma.
• To assess the performance of this signature compared to individual radiomic features and conventional imaging/clinical data.

Main Methods:

• Radiomic features were extracted from APT-related metric maps and structural MRI scans of 56 pediatric medulloblastoma patients.
• Feature selection was performed, and radiomic signatures were generated using the least absolute shrinkage and selection operator (LASSO) method.
• Receiver operating characteristic (ROC) analysis was used to evaluate the performance of individual features and the radiomic signature for risk stratification.

Main Results:

• The optimal individual radiomic feature achieved an area under the ROC curve (AUC) of 0.83 for risk stratification.
• The developed APT MRI-based radiomic signature demonstrated a higher AUC of 0.86 in distinguishing medulloblastoma risk groups.
• Incorporating structural MRI data or clinical factors did not improve the diagnostic performance of the APT-based radiomic signature.

Conclusions:

• An APT MRI-based radiomic signature is effective for differentiating pediatric medulloblastoma risk subgroups.
• This radiomic approach offers valuable, non-invasive insights for improving the clinical diagnosis and management of medulloblastoma.
• The findings suggest that APT MRI radiomics can serve as a powerful tool in the precision medicine approach to pediatric brain tumors.